ReClock 1.8.6.2

[James]

glad I'm not the only one getting crackles

Likely a different problem (XP vs. Vista/W7, KS vs. Wasapi).

 

[leeperry]

Likely a different problem (XP vs. Vista/W7, KS vs. Wasapi).

well yeah, but b55 works perfectly fine for me...and the problem seems to be buffer related in both cases somehow

you didn't change anything in the KS code AFAIK, so the problem is most likely upstream...anyway, I guess the XP/KS support didn't "break" suddenly all by magic >b58..

I will continue to slowly update, and hopefully be able to provide useful logs

PS: to be perfectly clear, I'm not getting loud "glitches" like when Reclock needs to catch up, I'm getting very soft crackles/pops like when I set the audio buffer too low.

 

[James]

PS: to be perfectly clear, I'm not getting loud "glitches" like when Reclock needs to catch up, I'm getting very soft crackles/pops like when I set the audio buffer too low.

Is this "audio only" mode or "normal" mode? What are your buffer settings?

 

[leeperry]

Is this "audio only" mode or "normal" mode? What are your buffer settings?

both A/V buffers are set to 500ms, and problem happens both w/ A and V files...very random, rewinding doesn't make it happen again...so I don't have much faith in being able to produce valuable logs

b55 doesn't exhibit this problem at all, but the SQ is pretty bad compared to b58...which I'm currently running, and everything looks fine so far :agree:

I think the problem started w/ b60(using your own SRC DLL)...but hopefully Jong will have valuable logs to show you.

 

[James]

both A/V buffers are set to 500ms, and problem happens both w/ A and V files...very random, rewinding doesn't make it happen again...so I don't have much faith in being able to produce valuable logs

b55 doesn't exhibit this problem at all, but the SQ is pretty bad compared to b58...which I'm currently running, and everything looks life so far :agree:

I think the problem started w/ b60(using your own SRC DLL)...but hopefully Jong will have valuable logs to show you.

And this problem is independent if the resampler is used or not?

 

[leeperry]

And this problem is independent if the resampler is used or not?

it would appear so, yes...it seems to be like some hiccup in the way the buffer is being filled and streamed

I also tried w/ the stock resampler btw...but I'll run more tests on my side, and try to come up w/ more "clues".

 

[James]

I also tried w/ the stock resampler btw...but I'll run more tests on my side, and try to come up w/ more "clues".

The stock resampler won't improve anything if it happens without...

 

[Bitmonster]

@James
Could you please explain the audio clock "samp" value that is displayed in the properties dialog, if AC3 bitstream passthrough is used? I have searched the forum but haven't found a suitable answer.

For PAL DVD it shows 32ms and this makes sense to me, since AC3 packets are 32ms long.
But for 23.976 fps and 24.000 fps it shows 11ms, and this sounds odd to me. Shouldn't it also be 32ms? Or does ReClock have the ability to "shift" the packets on the output, so they don't need to be aligned to 32ms? But if so, why doesn't it do it for 25.000 fps also?

 

[ashlar]

Yes.
......

Care to expand on this (64 bit), please?

 

[Bitmonster]

Care to expand on this (64 bit), please?

Have you read, what he already wrote on this topic?
http://forum.slysoft.com/showthread.php?t=23247

 

[Bitmonster]

@James
Could you please explain the audio clock "samp" value that is displayed in the properties dialog, if AC3 bitstream passthrough is used? I have searched the forum but haven't found a suitable answer.

For PAL DVD it shows 32ms and this makes sense to me, since AC3 packets are 32ms long.
But for 23.976 fps and 24.000 fps it shows 11ms, and this sounds odd to me. Shouldn't it also be 32ms? Or does ReClock have the ability to "shift" the packets on the output, so they don't need to be aligned to 32ms? But if so, why doesn't it do it for 25.000 fps also?

Ok, I can partly explain it myself.
The 11ms are actually DTS audio. Haven't seen it, because in bitstream mode ffdshow doesn't show the type of the stream. And it was new to me DTS uses 11ms packets (exactly 10.66666... ms?).

Well actually it should answer the question as a whole. So the "samp" is just the packet size?

 

[James]

So the "samp" is just the packet size?

Yes.
.............

 

[James]

Care to expand on this (64 bit), please?

ReClock is a major hack (just think about what it does for a moment) which simply isn't as easy to port to 64 bit as you might think. It is possible, but not just "changing a compiler option". It is a lot of work. This takes time. I don't have it. At the moment, benefits are minor. So I'm not doing it.

 

[yesgrey3]

Sorry for this off-topic post, but since I've mencioned it before, I think I can add just one more post...

If Microsoft had done only one very simple thing with W7, I would be on it right now, but they had decided not to do it, so probably soon I will get back to my XP (my W7 RC1 is already near the end).
What was that simple think, you may ask?
Just keeping as an option the exact Windows2000 GUI theme, like they did with XP. I simply don't understand why they had removed it.
W7 has a similar one, but it's not the same in several details, and those details are really painful...:bang:

As for the option of the GUI, I wish I knew. There are things that annoy me too in the Win7 GUI, I just had to compromise. For the power user Vista and Win7 is a step back, which is unfortunate.

I think I've found the solution for the several details that were annoying me.
Microsoft did not keep the exact Windows XP classic GUI, but someone else decided to recover those little details that are missing in W7 and are a real pain to a lot of users.
With this little program we are able to get the XP windows explorer in all its glory without the window default size bug that always made me mad about it...
If you want to try it, go here, it's free.

 

[ashlar]

ReClock is a major hack (just think about what it does for a moment) which simply isn't as easy to port to 64 bit as you might think. It is possible, but not just "changing a compiler option". It is a lot of work. This takes time. I don't have it. At the moment, benefits are minor. So I'm not doing it.

Ok. Thanks for the explanation. I hope in the long run you'll find the time. But I appreciate the enormous amount of work you do on Reclock, I really do.
Cheers!

 

[Bitmonster]

Yes.
.............

Ok, then I have a question, that I really don't understand.

I have a refresh rate timing for 23.976 fps that is perfectly "calibrated" (I have written a tool to calculate such fine grained timings). With this timing ReClock shows less then 1 ms audio drift per hour with DTS bitstream passthrough. So nearly no drift and ReClock never has to drop/repeat a packet in a movie.

If I now use AC3 passthrough with the same timing (and a movie with 23.976 fps of course), ReClock shows a drift of about 30ms per hour.

How is this possible?

 

[James]

Ok, then I have a question, that I really don't understand.

I have a refresh rate timing for 23.976 fps that is perfectly "calibrated" (I have written a tool to calculate such fine grained timings). With this timing ReClock shows less then 1 ms audio drift per hour with DTS bitstream passthrough. So nearly no drift and ReClock never has to drop/repeat a packet in a movie.

If I now use AC3 passthrough with the same timing (and a movie with 23.976 fps of course), ReClock shows a drift of about 30ms per hour.

How is this possible?

What should I say? What I said many times: don't use ReClock with bitstreaming. Let it do its job, or don't use it.

 

[leeperry]

The stock resampler won't improve anything if it happens without...

I've listened to music(16/44.1 FLAC in bit-perfect KS) w/ b55/b58 for many hours, no weird noises at the end of songs...but they're back on b62, soft crackles/pops again(and they don't reappear if I rewind)...they seem to happen when the buffer gets emptied/flushed/reopened by the next file in the playlist(w/ MPC/madflac/pass-through ffdshow).

I'll try to roll back to b61/b60 now..

PS: b61 also does it, currently trying b60

PPS: b60 appears to work perfectly fine w/ both A and V files, I'll keep testing

 

[FoLLgoTT]

What should I say? What I said many times: don't use ReClock with bitstreaming. Let it do its job, or don't use it.

But one part of ReClock's job is to drop/repeat samples of the bitstream if audio and video run out of sync. I mean, in many situations it is better not to decode AC3/DTS/TrueHD in software and let the work be done by an external receiver.

Will there be a fix for bitstreaming formats other than AC3?

 

[leeperry]

I dunno for Jong, but I can confirm that audio is completely glitch-free in b60 for me(whatever audio files or videos using SRC)...so something seems to have happened in b61

But one part of ReClock's job is to drop/repeat samples of the bitstream if audio and video run out of sync. I mean, in many situations it is better not to decode AC3/DTS/TrueHD in software and let the work be done by an external receiver.

Will there be a fix for bitstreaming formats other than AC3?

Reclock needs to resample the audio at a variable sample rate to match the actual hardware video refresh rate...this cannot be done in bitstream.